Perform vector similarity search on a collection. Args: collection_name: Name of the collection to search vector: Query vector vector_field: Field containing vectors to search limit: Maximum number of results output_fields: Fields to include in results metric_type: Distance metric (COSINE, L2, IP...
AI agents call milvus_vector_search to retrieve information from MCP Server for Milvus without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves data through vector similarity search without any side effects or state changes. It accepts search parameters and returns results, which is characteristic of a Read operation. The lack of any modify, delete, create, or execute keywords, combined with the purely query-based nature of vector search, confirms this is a retrieval operation with minimal risk if misused by an AI agent.
From the tool's definition Tool description states 'Perform vector similarity search on a collection' with parameters for querying (collection_name, vector, limit, output_fields, metric_type, filter_expr, radius, range_filter).
Documented attack patterns abuse exactly the kind of access milvus_vector_search gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and MCP Server for Milvus, and nothing reaches the server without passing your rules. This is the rule we recommend for milvus_vector_search:
{
"version": "1",
"default": "deny",
"tools": {
"milvus_vector_search": {}
}
}milvus_vector_search is read-only, so it stays allowed — but everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Perform vector similarity search on a collection. Args: collection_name: Name of the collection to search vector: Query vector vector_field: Field containing vectors to search limit: Maximum number of results output_fields: Fields to include in results metric_type: Distance metric (COSINE, L2, IP) filter_expr: Optional filter expression radius: Optional lower bound for range search range_filter: Optional upper bound for range search. It is categorised as a Read tool in the MCP Server for Milvus MCP Server, which means it retrieves data without modifying state.
Register the MCP Server for Milvus MCP server in PolicyLayer and add a rule for milvus_vector_search: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches MCP Server for Milvus. Nothing to install.
milvus_vector_search is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the milvus_vector_search rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for milvus_vector_search. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
milvus_vector_search is provided by the MCP Server for Milvus MCP server (zilliztech/mcp-server-milvus). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Deterministic rules across all 14 MCP Server for Milvus tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.
Free to start. No card required.
14 MCP Server for Milvus tools catalogued and risk-classified — across an index of 42,500+ MCP servers.